R
Rob
Hi everyone,
I need some help / advice.
I'm trying to figure out how I can determine an equation that will describe
my data. I am measuring product performance (PP) vs. time (T). In general
the products can behave in at least 5 ways as they age:
1. an initial performance improvement followed by a slow decline.
2. no change with time.
3. a slow decrease in performance until a threshold is reached, then a rapid
decrease in performance.
4. a straight-line decrease in performance.
5. a rapid initial decrease in performance, followed by a levelling off.
It is also possible to have combinations of these if several components
degrade at different rates, or degradation of each component is triggered by
different events.
The data has already been collected, and plotted. I would like to be able
to determine an equation that will describe the data, so that PP can be
predicted at any given T. I would also like to be able to determine the
level of confidence, etc. for the equation.
I've tried using LOGEST and LINEST but they don't really produce very good
results with this data. I've tried using Trendlines but it is sometimes
difficult to find a line that matches the data well, and it doesn't provide
any supporting statistics. Is there a tutorial that could help me get a
better understanding? Are there plug-ins available that would help?
Thank you,
Rob
I need some help / advice.
I'm trying to figure out how I can determine an equation that will describe
my data. I am measuring product performance (PP) vs. time (T). In general
the products can behave in at least 5 ways as they age:
1. an initial performance improvement followed by a slow decline.
2. no change with time.
3. a slow decrease in performance until a threshold is reached, then a rapid
decrease in performance.
4. a straight-line decrease in performance.
5. a rapid initial decrease in performance, followed by a levelling off.
It is also possible to have combinations of these if several components
degrade at different rates, or degradation of each component is triggered by
different events.
The data has already been collected, and plotted. I would like to be able
to determine an equation that will describe the data, so that PP can be
predicted at any given T. I would also like to be able to determine the
level of confidence, etc. for the equation.
I've tried using LOGEST and LINEST but they don't really produce very good
results with this data. I've tried using Trendlines but it is sometimes
difficult to find a line that matches the data well, and it doesn't provide
any supporting statistics. Is there a tutorial that could help me get a
better understanding? Are there plug-ins available that would help?
Thank you,
Rob